Re: pgsql: Add a regression test for allow_system_table_mods

On Fri, Nov 29, 2019 at 11:09:06AM -0500, Tom Lane wrote:
> I agree that this is somewhat pointless in the case of an "unsafe" test.
> But using a different name isn't going to invalidate the test case,
> so there's not really a reason to not follow the convention.  And
> trying to have an exception for unsafe_tests seems like a lot more
> trouble than it'd be worth.

Yeah, I agree that it would be just more simple to make the unsafe
test adopt the convention and be done with it.  Could you fix it
please?  longfin and mantid (as well as anybody running the unsafe
tests by themselves) are complaining for three days now.
--
Michael